Automated Solid Identification in Sewer Systems

2020 Research Internship Project

---

Faculty Name

Dimitri Androutsos

Project Title

Automated Solid Identification in Sewer Systems

Project Description

Through image-based sensors and developed algorithms, the student will work on developing an automated in-sewer image/video capture, transfer, and processing system. To consume battery power, the image-based sensor will only transfer the image/video upon detection of material. The material to be identified is gross solids, including consumer products like baby wipes, paper towel, toilet paper, and more, as well as other common solids found in sewers like twigs and branches. The algorithm will use training data sets and associative memories so that products can automatically be classified into subcategories of consumer products.

Student Responsibility

The projet will require the student to use MATLAB to process the digital data captured by the iaging system. The data collected must be processed into such a form that is readable by MATLAB or similar software. Additionally, to apply image/video pre-processing on the dataset, the following steps will be taken: 1. Read image (important to establish a base size for images fed into the AI algorithms) 2. Resize image 3. Remove noise (denoise) 4. Contrast enhancement 5. Segmentation 6. Morphology (smoothing edges)

Specific Requirements

- knowledge of MATLAB - strong understanding of signals and systems - understanding or experience in working with images is a plus